Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Kögler, Kevin, Shevchenko, Alexander, Hassani, Hamed, Mondelli, Marco
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929237129166848
author Kögler, Kevin
Shevchenko, Alexander
Hassani, Hamed
Mondelli, Marco
author_facet Kögler, Kevin
Shevchenko, Alexander
Hassani, Hamed
Mondelli, Marco
contents Autoencoders are a prominent model in many empirical branches of machine learning and lossy data compression. However, basic theoretical questions remain unanswered even in a shallow two-layer setting. In particular, to what degree does a shallow autoencoder capture the structure of the underlying data distribution? For the prototypical case of the 1-bit compression of sparse Gaussian data, we prove that gradient descent converges to a solution that completely disregards the sparse structure of the input. Namely, the performance of the algorithm is the same as if it was compressing a Gaussian source - with no sparsity. For general data distributions, we give evidence of a phase transition phenomenon in the shape of the gradient descent minimizer, as a function of the data sparsity: below the critical sparsity level, the minimizer is a rotation taken uniformly at random (just like in the compression of non-sparse data); above the critical sparsity, the minimizer is the identity (up to a permutation). Finally, by exploiting a connection with approximate message passing algorithms, we show how to improve upon Gaussian performance for the compression of sparse data: adding a denoising function to a shallow architecture already reduces the loss provably, and a suitable multi-layer decoder leads to a further improvement. We validate our findings on image datasets, such as CIFAR-10 and MNIST.
format Preprint
id arxiv_https___arxiv_org_abs_2402_05013
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth
Kögler, Kevin
Shevchenko, Alexander
Hassani, Hamed
Mondelli, Marco
Machine Learning
Information Theory
Autoencoders are a prominent model in many empirical branches of machine learning and lossy data compression. However, basic theoretical questions remain unanswered even in a shallow two-layer setting. In particular, to what degree does a shallow autoencoder capture the structure of the underlying data distribution? For the prototypical case of the 1-bit compression of sparse Gaussian data, we prove that gradient descent converges to a solution that completely disregards the sparse structure of the input. Namely, the performance of the algorithm is the same as if it was compressing a Gaussian source - with no sparsity. For general data distributions, we give evidence of a phase transition phenomenon in the shape of the gradient descent minimizer, as a function of the data sparsity: below the critical sparsity level, the minimizer is a rotation taken uniformly at random (just like in the compression of non-sparse data); above the critical sparsity, the minimizer is the identity (up to a permutation). Finally, by exploiting a connection with approximate message passing algorithms, we show how to improve upon Gaussian performance for the compression of sparse data: adding a denoising function to a shallow architecture already reduces the loss provably, and a suitable multi-layer decoder leads to a further improvement. We validate our findings on image datasets, such as CIFAR-10 and MNIST.
title Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth
topic Machine Learning
Information Theory
url https://arxiv.org/abs/2402.05013